Crystal structure of bis (9-methylhypoxanthine) silver (I) perchlorate

Belander-Gariepy, Beauchamp / Bis (9-met hy ¡hypoxanthine) silver (I) Perchlorate Monohydrate ... Perchlorate Monohydrate, a Model Complex for the Si...
0 downloads 0 Views 519KB Size
Belander-Gariepy, Beauchamp

/

Bis (9-met hylhypoxant hine)siloer ( I ) Perchlorate Monohydrate

3461

Crystal Structure of Bis( 9-methylhypoxanthine)silver(I) Perchlorate Monohydrate, a Model Complex for the SilverPoly( I) System Francine Belanger-Gariepy and Andre L. Beauchamp* Contribution from the Dfpartement de Chimie. Universitf de Montrfal. C.P. 6210, Succ. A, Montrial. QuCbec. Canada H 3 C 3 V l . Received October 11, 1979 Abstract: Crystals of the title compound belong to space group Pi with the following cell parameters: a = 9.271 (3) A, b = 10.187 (7) A, c = 19.124 (7) A, a = 91.17 (4)O, /3 = 94.08 (3)O, and y = 93.78 (4)O. The structure consists of uncoordinated C104- ions and planar [(9-methylhypoxanthine)2Ag]+ cations in which silver is linearly bound to two ligand N(7) Atoms. The ligands are syn with respect to N(7)-Ag-N(7). Carbonyl oxygen atoms O(6) are not bound to silver, but they are cross-linked via hydrogen bonding by a water molecule. Such cations associate into dimers by hydrogen-bond formation between the water molecule and a H ( 1) proton belonging to the other [(9-methylhypoxanthine)zAg]+ cation in the asymmetric unit. The present structure provides a reasonable structural modql for the 2:l poly(1)-Ag complex observed by Shin and Eichhorn a t low pH and it suggests assumptions for the structure of the 4:4 complex reported for the same system at higher pH.

Silver reacts with DNA without loss of the helical structure.’ This behavior is connected with the preference of silver for linear two cootdination and its affinity for nitrogen donors on the bases. Propensity to cross-link strands and stabilize multiple helices also extends to synthetic polynucleotides. Interesting results have recently been reported on the polyriboinosinic acid (poly(1)) system by Shin and Eichhorn.2 The great affinity of Ag(1) for hypoxanthine residues of poly(1) reported earlier3 was confirmed by this more recent work and two types of complexes were identified. At pH 4, reaction takes place with two residues without exchange of protons, forming a double helix formulated as poly(Ag-12). On the other hand, at pH 8, silver reacts according to a 1: 1 ratio and displaces one H+ ion. On the basis of X-ray fiber diagrams, formation of a four-stranded helix poly(Ag4-14) was proposed. The 1:2 complex with 9-methylhypoxanthine described in the present paper was prepared under slightly acidic conditions. A crystallographic study was undertaken because the compound was stoichiometrically, and presumably structurally, related to poly(Ag-It). This structure also suggests possible arrangements of silver atoms and hypoxanthine moieties in the poly(Ag4-14) complex observed at higher pH.

Experimental Section Preparation. 9-Methylhypoxanthine (0.10 g) was dissolved in water (25 mL) and 7.0 mL of 0.090 M AgC104 in 0.66 M HC104 was added. The pH was raised to 4.0 with 1 M NaOH (-1 mL). The cloudy mixture was gently heated and the resulting clear solution was allowed to cool slowly to room temperature. After a few hours, crystals separated from the viscous solution. They were isolated and dried in air. Anal. Calcd: Ag, 20.52; CI, 6.75; C, 27.42; H, 2.68; N, 21.32. Found: Ag, 20.48; CI, 6.77; C, 26.85; H, 2.68; N, 20.76. Crystal Data. (C&N40)2A CIOgH20, mol wt 525.61, triclinic, space group Pi, a = 9.271 (3) b = 10.187 (7) A, c = 19.124 (7) A, a = 91.17 (4)O, /3 = 94.08 (3)O, y = 93.78 (4)O, V = 1797.2 A3, Z = 4 formula units per cell, dc = 1.942 g ~ m - p~ =, 13.0 cm-’, A = 0.710 69 A (graphite monochromator), t = 22 O C . Crystallographic Measurements and Structure Resolution. The crystals were elongated plates of a rather poor quality. Attempts to improve crystal growth by changing concentrations, pH, or cooling rate were unsuccessful. One of the best specimens was selected for space-group determination and used to collect a set of intensity data with which the structure was solved. The final R factor (0.068) was considered to be too high and a second set of data was collected with a new crystal. This set refined slightly better and it is used in this paper. The second crystal had the folloEing dimensions beiw_eenthe indicated pairs of fases: 0.03 (001 and OOI), 0.35 ( 1 10 and 1 IO), and 0.20 mm (100 and 100). The triclinic Laue symmetry observed gn precession photographs was consistent with space groups P1 and P1. Accurate cell parameters

1,

0002-7863/80/1502-3461$01 .OO/O

were derived from least-squares refinement of the setting angles for 25 reflections centered in the counter aperture of an Enraf-Nonius CAD-4 diffractometer. Intensity data were collected using the w/26 scan technique. A fixed slit 4.0 mm wide was used and the scan range (0) was (0.75 0.35 tan 6)O extended 25% on either side for background. A scan time of 120 s was imposed, which corresponds to scan speeds of --0.50°min-I. Prescans a t 6.7’ min-l, with background counts, were done A d all reflections having 1/a(1)< 1 were labeled “weak” and not remeasured. Three standard reflections were used as a check on instrument and crystal stabilhy. They showed random fluctuations within f 4 % about their respective means. Crystal orientation was checked every 100 reflections by recentering three axial reflections. No significant motion of the crystal was detected during data collection. A total of 4678 independent reflections (hkl, zkl, hi& and zzl, 26 < 45’) were measured using Enraf-Nonius option FLAT, a procedure to measure intensity at the J/ position where absorption is minimum. Net intensities, I , and their esd’s, a(1),wme calculated as described previ~usly.~ A set of 2985 reflections with 1 3 2 4 1 ) was retained for the structuie determination. An absorption correction was applied (Gaussian integration, grid I O X 10 X IO); the transmission factors were in the range 0.78-0.96. Finally, the data were corrected for LP. The structure was solved by the heavy-atom method and refined by full-matrix least squares initially-and by block-diagonal least squares in the latest stages. In group PI, the asymmetric unit contains two nonequivalent silver atoms, which were located from a Patterson synthesis. The positions of the remaining nonhydrogen atoms were determined from subsequent Fourier maps. Refinement of positional and isotropic thermal parameters of all nonhydrogen atoms using unit weightsconverged t o R = 21 IFoI - lFcl l/ZIFol = 0.142. Hydrogen atoms on N( l), C(2). and C(8) of each ring were then fixed at their calculated positions with isotropic B’s of 6.0 A*. Those of the methyl groups and water molecules were not located. Anisotropic refinement on all nonhydrogen atoms reduced R to 0.07 1. At this point, a Fourier map clearly showed that c104- ion no. 2 is disordered: the abnormally high temperature factor of O(8) suggested partial site occupancy, whereas a spurious peak, 0 ( 8 ‘ ) , remained at a suitable distance from Cl(2) to correspond to a bonded oxygen atom. Also, dissymmetric electron density contours suggested the presence of overlapping components for thejest of this ion. Although refinement of a PI structure in group P1 could lead to apparent disorder, the absence of similar effects elsewhere in the cell ruled out such an explanation. Hence, attempts were made to describe the electron de_nsity around a ( 2 ) in terms of two overlapping c104- ions in group PI. Occupancy factors were first refined for O(8)and 0 ( 8 ’ ) , the only resolved components. The refined values normalized to unity (0.60 and 0.40, respectively) were then assumed to hold for their respective ions. Various sets of positions for the tetrahedra were tried, using 0 ( 8 ) and O(8’) as pivots, but no real improvement resulted in the R factor. Since a large amount of computer time had already been spent just to improve this structural detail of limited interest, anisotropic refinement was terminated using individual positions for 0 ( 8 ) and O(8’)only. In the

+

0 1980 American Chemical Society

Journal of the American Chemical Society

3462

/

102:lO

/

May 7 , 1980

Table I. Refined Coordinates of Bis(9-methy1hypoxanthine)silver(1) Perchlorate Monohvdrate (X103. Ag, CI X104) atom

13

N43

Figure 1. The asymmetric unit of [(9-methylhypoxanthine)2Ag]C104. H20. The first and second digits in symbols for ring atoms correspond to ring number and atom number in the ring, respectively. O( 11) and O( 10) are water molecules.Dashed lines represent H bonds. Ellipsoids correspond to 50% probability. Hydrogen atoms at N(il), C(i2), and C(i8) are not represented. Y

I

/n\

/\ Figure 2. Interatomic distances and bond angles. For ligand molecules, values averaged over the four independent ligands are given. The remaining values have been averaged over the two nearly centrosymmetric halves of the asymmetric unit. Standard deviation on bond lengths are 0.01-0.02 8, except for Ag-N, 0.01 A, and on bond angles 1' except for the angles at Ag, 0.4'.

final stage of the refinement, individual weights based on counting statistics were used. The final R factor was 0.062, while R, = z(w[lFoI - ~ F c ) ] 2 / ~ ~ I F o 1 2was0.080. )'/2 The refined coordinates are listed in Table I. Anisotropic temperature factors and hydrogen coordinates are part of the supplementary material. The form factors used were those of Cromer and Waber,5 except for hydrogen6 Anomalous scattering termsf' andf" were used for Ag and Cl.'

Description of the Structure Besides two C104- ions, the asymmetric unit includes two

X

).'

Z

4323( 1) 5454( 1) 677(1) 120(1) 350( 1) 882(1) 765(1) 34(1) 257(1) 949( 1) 795(1) 27(1) 217(1) 938( 1) 719(1) 115(1) 272( 1) 846( 1) 628(1) 208(1) 366( 1) 771(1) 598(1) 215(1) 412(1) 789(1) 518(1) 291(1) 493( 1) 737( 1) 576( 1) 280( 1) 399( 1) 689( 1) 637( 1) 234( 1) 325(1) 716(1) 726( 1) 132(1) 251(1) 812(1) 808(2) 50(2) 164(2) 869(2) 313(4) 44 1O( 5 ) 56(2) 163(1) -39( 1) -42(2) 306( 1) 431(2) 515(2) 525(2) 404(4) 364(1) 673(1)

3256( 1) 1881(1) 95(1) 651(1) 457( 1) -108(1) -1(1) 720(1) 552( 1) -194(1) -35(1) 704( 1) 575(1) -201(1) 31(1) 613(1) 495( 1) -1 15( 1) 129( 1) 544( 1) 395(1) -26( 1) 167(1) 559( 1) 373( 1) -18(1) 250( 1) 507( 1) 292(1) 62(1) 174( 1) 462( 1) 338(1) 42(1) 104(1) 486( 1) 398(1) -3(1) 18(1) 577( 1) 495(1) -99( 1) -78(2) 624( 1) 585(1) -167(1) 7579(4) 2037(5) 879(2) 698(2) 776( 1) 675(1) 138(2) 241(2) 236(2) 80(2) 363(3) 398(1) 173(1)

551(1) -5350(1) - 134( 1) -61(1) -346( 1) -412(1) -114(1) -20(1) -366( 1) -451( 1) -52( 1) 47(1) -430( 1) -518(1) -5(1) 7 ~ -477( 1) -547( 1) -19(1) 36(1) -463(1) -512(1) -90( 1) -39( 1) -392(1) -438( 1) -111(1) -77( 1) -371(1) -401(1) 41(1) 81(1) -523(1) -559( 1) 92(1) 142(1) -57 1(1) -620(1) 66( 1) 140( 1) -547( 1) -616(1) 108( 1) 199( 1) -589( 1) -676( 1) -2404(2) 2628(2) -271( 1) -218(1) -180(1) -290(1) 283(1) 201(1) 322( 1) 245( 1) 271(1) -205(1) -273(1)

nonequivalent AgL2+ complex cations (where L = 9-methylhypoxanthine) and a pair of water molecules O(10) and O( 11) (Figure 1). Interatomic distances and bond angles are schematically represented in Figure 2. Both silver atoms are similarly bonded to two imidazolic N(7) atoms, the metal showing the usual linear two coordination. Departure from linearity is significant (N-Ag-N = 173.1 (4)O), but within a range commonly found for silver. The Ag-N bonds (2.13 (1) A) are somewhat longer than observed in the 1-methylthymine complex (2.081 A): but shorter than those of compounds with 1-methylcytosine (2.225 A),9 9-

)

Belander-Gariepy, Beauchamp

/ Bis( 9-methylhypoxanthine)siluer(I)Perchlorate Monohydrate

methyladenine (2.16 A),1oand adeninium cation (2.194 and 2.154 A ) . l l The shortest distance to a carbonyl oxygen is 3.37 A, which is well beyond the 3.0-A limit considered by many authors’for bonding interacti0ns.l Interestingly, the ligands in each AgL2+ cation are coplanar and a water molecule forming moderately strong H bonds to carbonyl groups (O--O = 2.78 A) helps to keep the six-membered rings syn with respect to N-Ag-N. In turn, that water molecule is tightly attached to the adjacent AgL2+ cation via a strong N( 1)-H( 1)-.0 hydrogen bond .(N-O = 2.78 A). Two such water molecules acting as identical three-way links hold two AgL2+ ions in a nearly planar centrosymmetric unit. The maximum tilt of any individual ligand with respect to the average plane of the “diaquo dimer” is only 10’. The four nonequivalent ligands have the same geometry within experimental errors.12 Bond lengths and angles agree well with those of similar m o l e ~ u l e s . ~The ~ J C(2)-N( ~ 1)-C(6) angles average 126’, which confirms the presence of a hydrogen atom on N ( l), that angle being normally smaller when N ( l ) is not pr0t0nated.l~The purine rings are planar within 1.5a (0.015 A) and the atom-to-plane distances for the exocyclic groups are significant but not unusual: 0(46), 0.1 1 A; methyl groups, 0.04-0.08 A; Ag, 0.04-0.12 A; a = 0.01 A. Perchlorate ion 1 shows a normal geomk.try:I2 the C1-0 distances range from 1.37 (2) to 1.44 (2) 8, (av 1.39 A) and the 0-C1-0 angles from 105 (1) to 115 (1)’ (av 109’). The spread in bond lengths and angles for ion 2 is due to incomplete resolution of the disordered components (vide supra). The packing pattern (Figure 3) shows the parallel stacking of “diaquo dimers”. Columns of C104- ions occupy channels parallel to the a axis and they interact only moderately with the rest of the structure. Ions 2 act as bridging units between Ag( 1) of one dimer and Ag(2) of the next dimer along c. One of the contacts (Ag( 1)-0(6) = 2.94 (1) A) is just within the 3.0-A limit, while the other (Ag(2)-0(7) = 2.78 (2) A) is shorter, but still much longer than found in the adeninium complex (2.64 A).11Ion 1 forms weak H bonds with the two N(1)-H( 1) groups not already bonded to water: N(41)-0( 1) = 3.06 A; N(21)-0(2) = 3.10 A; N(21)-*0(3) 2.98 A; = 0.01 A.

Discussion Silver binding to N(7) is not surprising. Metal ions usually bind to that site in N(9)-blocked hypoxanthine or guanine rings.I4J6 In many of those structures, other ligands in the coordination sphere were found to form intramolecular H bonds to the adjacent carbonyl oxygen O(6). However, no such ligands are available with linearly coordinated silver, with the consequence that the basic character of O(6) must be satisfied differently. This is achieved by water molecules forming H bonds and “cross-linking” the two purine residues of the AgL2+ cation. Shin and Eichhorn2 have proposed N(7)-Ag-N(7) bonding for the poly(Ag-12) double-helix complex formed at pH 4 and this assumption is fully confirmed by the present crystal structure. Ideally, free hypoxanthine rings could be positioned in syn (Figure 2) or anti configurations with respect to N-

I

I1

I1 I

3463

Figure 3. Packing diagram of [ (9-methylhypoxanthine)~Ag]CIO~.H~O. Dashed lines correspond to H bonds.

Ag-N. The syn structure was observed in the present compound and is proposed as a probable model for poly(Ag-12). However, structure I cannot be ruled out even if stabilization by water “cross-linking” would be lost, because that particular orientation of purine rings may have been forced by other factors in the more constrained poly(1)-Ag system. The ligand hydrogen H ( 1) is weakly acidic and N( 1) becomes available for complexation at higher pH. Structural information on silver compounds with deprotonated hypoxanthine rings is still missing, but one of the reasonable assumptions that can be put forward would be closely related to the “diaquo dimer” structure observed here. Assuming that the two water molecules and four H ( 1) protons are removed and that the two AgL2+ cations are shifted with respect to each other in order to position N(11) and N(21) in front of N(41) and N(31), respectively, reaction of two silver atoms with the four available N(i1) donors would lead to the planar tetramer 111. Such a unit would have a 1:l stoichiometry with substitution of one H+ per Ag+ ion and all the donors bound to silver would be nitrogen atoms, for which silver is known to show a greater affinity than for oxygen. This model is particularly attractive in connection with the evidence provided by Shin and Eichhorn2 for a four-stranded helix in the poly(1)-Ag complex formed at high pH. They have proposed structural unit I1 in which N( 1) and O(6) are coordinated to silver, leaving N(7) free. This model is geometrically reasonable, but it requires coordination with an oxygen donor when a nitrogen donor N(7) is available. Assuming that reaction is forced to take place with O(6) and N( l), the binding curve should not level at a 1:1 Ag:residue ratio, since free N(7) atoms would be expected to accept mpre silver ions. On the other hand, structure I11 would make efficient use of both N( 1) and N(7) on all four ligands, providing silver with a linear coordination of nitrogen donors only. Extra stabilization could result from water molecules filling the central channel and cross-linking carbonyl groups of adjacent strands. Other arrangements with different relative orientations of the bases are also geometrically and chemically reasonable. For instance, interchanging N(7) and N ( l ) of two hypoxanthine moieties on opposite corners of the square would lead to a symmetric unit with a fourfold axis, in which all Ag atoms would have the same N( l)-Ag-N(7) coordination. Arrangement I11 has been preferred because the structure of Figure 2 could be its precursor, thus providing a convenient mechanism for the equilibrium between poly(Ag-12) and poly(Ag4-14) noted for intermediate pH values by Shin and Eichhorn.2 Supplementary Material Available: Temperature factors of nonhydrogen atoms, calculated coordinates of hydrogen atoms, individual interatomic distances and bond angles, least-squares planes calculated through purine rings, and table of calculated and observed structure factors with standard deviations (X 10) (24 pages). Ordering information is given on any current masthead page.

3464

Journal of the American Chemical Society

References and Notes (1) Eichhorn, G. L. “Inorganic Biochemistry”; Elsevier: Amsterdam, 1973; Chapter 34. (2)Shin, Y. A,; Eichhorn, G. L. 6iopolymers 1980, 19, 539. (3)Eichhorn, G. L.; Butzow, J. J.: Clark, P.; Tarien, E. Biopolymers 1967, 5, 283. (4)Authier-Martin, M.; Hubert, J.; Rivest. R.; Beauchamp. A. L. Acta Crystallog., Sect. 8 1978, 34,273. (5) Cromer, D. T.; Waber, J. T. Acta Crystallogr. 1965, 78, 104. (6)Stewart, R. F.;Davidson, E. R.; Simpson, W. T. J. Chem. fhys. 1965, 42, 3175. (7)Cromer, D. T. Acta Crystallogr. 1965, 18, 17. (8)Guay, F.; Beauchamp, A. L. J. Am. Chem. Soc. 1979, 701, 6260. (9) Kistenmacher, T. J.; Rossi, M.; Marzilli, L. G. lnorg. Chem., 1979, 78, 240.

(10)Gagnon, C.; Beauchamp, A. L. Acta Crystallogr., Sect. 6 1977, 33,

/

102:lO

/

May 7 , 1980

1448. (11) Gagnon, C.; Hubert, J.; Rivest, R.; Beauchamp, A. L. lnorg. Chem. 1977, 16. 2469. (12) Individual bond lengths and angles, best plane equations, and atomto-phne distances can be found in the supplementary material. (13) TheWalt, U.;Bugg, C. E.; Marsh, R. E. Acta Crystallogr., Sect. 8 1970, 26, 1089.Munns, A. R. I.;Tollin, P. /bid. 1970, 26, 1101.Rao, S.T.; Sundaralingam, M. J. Am. Chem. Soc. 1969, 91, 1210. (14)Adti. K. Bull. Chem. Soc. Jin. 1975, 48.1260.Sletten. E. Acta Crystallog., Sect. 6 1974, 30, 1961.Sletten, E.; Kaale, R. bid. 1977, 33, 158. (15)Singh, C. Acta Crystallogr. 1965, 79, 861. (16)Hodgson, D. J. h o g . lnorg. Chem. 1977, 23,211.Melanson, R.; Rochon, F. D. Acta Cfystallogf., Sect. 6 1978, 34, 3594. Kistenmacher, T. J.; Chiang, C. C.; Chalipoyil, P.;Marzilli, L. G. J. Am. Chem. Soc.1979, 707, 1143.Behrens, N. B.; Cartwright. E. A,; Goodgame, D. M. L.; Skapski, A. C. lmrg. Chim. Acta 1978, 31,L471.Canty, A. J.; Tobias, R. S.:Chaichit, N.; Gatehouse, B. M.. private communication.

Sensitivity of the Cr-Cr Quadruple Bond to Axial Interactions in Dichromium( 11) Compounds F. Albert Cotton,* William H. Ilsley, and Wolfgang Kaim Contribution from the Department of Chemistry, Texas A & M University, College Station, Texas 77843. Received November 13, 1979

Abstract: X-ray crystallographic studies have been made on five Cr2[RNC(O)R’]pL2-type compounds, R = C6H5, 2,6(CH3)2CsH3; 4-(MezN)C&; R‘ = CH3, NHPh; L = THF, NC5H5, to determine the response of the Cr-Cr bond to the introduction and variation of axial ligands. The structures reported here are for the following compounds, where the numbers in parentheses are the lengths in 8, of the Cr-Cr bond: Cr2[(4-Me*N-C&4)NC(O)CH3]pTHF (1) (2.006 (2)); Cr2[(2,6xylyl)NC(O)CH3]4.3THF (2) (2.221 (3)); C ~ ~ [ ( ~ , ~ - X ~ I ~ ~ ) N C ( O ) C H ~ ] (3) ~ -(2.023 T H F .( IC) ) ;~ HCrz[PhNC~CH~ (NHPh)0]4.2THF.CsH14 (4) (2.246 (2)); Cr2[(2,6-xylyl)NC(O)CH3]4.3NC5Hs (5) (2.354 (5)). The Cr-Cr bond length is shown to be a sensitive function of both the number of axial ligands present and the donor ability of the axial ligands. Each T H F molecule in the axial position is shown to cause an -0.18-A increase in the length of the Cr-Cr bond. The Cr-Cr bond length is also seen to increase by about 0.1 2 A upon replacing two axial T H F ligands by two pyridine ligands. The basic crystallographic data on the five compounds are as follows. 1: space group, C2/c;a = 21.254 (4) A, b = 10.990 (2) A, c = 20.733 (4) A, fl = 110.64 (2)’. Z = 4. 2: Pbcn;a = 16.775 (2) A, b = 19.736 (3) A, c = 19.739 (2) A, Z = 4. 3 Pbcn; a = 16.806 (2) A, b = 19.31 1 (3) A, c = 14.887 (3) A, Z = 4.4: Pccn;a = 24.813 (5) A, b = 13.408 (3) A, c = 18.631 (3) A, Z = 4.5: P21/n; a = 17.374 (4) A, b = 12.836 (4) A, c = 23.603 (6) A, Z = 4. These structures are then compared with those of some 30 other quadruply bonded dichromium compounds with bridging ligand chains of the types 0 - C - 0 , N-C-0, N-C-N, N-N-N, CC-0, and C-P-C, and certain trends in the Cr-Cr distance as these ligand types are changed are noted.

Introduction The chemistry of compounds containing quadruple metalmetal bonds has developed rapidly since the recognition of the existence of the first such bond.’ With several of the elements that have shown a consistent tendency to form such bonds, particularly molybdenum and rhenium, the structural patterns are relatively simple. In these cases, the central M o ~ and ~+ Re26+ units appear to be structurally rigid and the observed variation in the metal-metal bond distance is relatively small. For example, Mo-Mo quadruple bond lengths range from 2.037 (3) A in M ~ ~ [ ( C S N H ~ ) N C ( O ) to CH 2.183 ~ ] ~(2) ~ A in M o ~ ( C ~ H and S ) ~the ~ known range of Re-Re quadruple bonds is from 2.178 (1) A in Re2Meg2- to about 2.24A in several Re2C1g2- s a h SWith chromium, however, the situation is more complicated. Quadruply bonded dichromium(IT) compounds exhibit a unique variability of the Cr-Cr bond length, covering a range of more than 0.7A. The shortest distance, 1.828(2) A, is found in Cr2[5-methyl-2-methoxyphenyl]4,and the longest, 2.541 (1) A, is found in Cr2(02CCF3)4(Et20)2.7Apart from the species [Cr2Rs14- (R = CH3 or ‘/*C4H@ and Cr2(C3H~)4~ which contain nonbridging ligands, the numerous examples of Cr2(X-Y-Z)4 compounds can be divided into two distinctly different groups: ( I ) First there are the tetracarboxylato compounds which have Cr-Cr distances ranging from 2.283 0002-7863/80/ 1502-3464$Ol .00/0

(2)to 2.541 (1) the related [Cr2(CO3)4l4- ion has a slightly shorter distance, 2.214(1) All of these molecules have been found to be either disolvated, in which the solvent molecules, L, occupy the axial positions, Cr2(02CR)4L2, or to have a structure in which the Cr2(02CR)4 units are arranged in infinite chains with intermolecular 0.-Cr axial distances of 2.22-2.44A.7 (2)Then there are the compounds with “supershort” bonds,12 Le., dcr-cr5 1.90A. These can generally be characterized as having ligands derived from weak acids and usually have the axial positions blocked as a result of intramolecular steric hindrance.I3 The rather long metal-metal distances in the carboxylato series have provoked considerable controversy concerning the Cr-Cr bond order.14$’sOne early report even claimed that in Cr2(02CCH3)4(H20)2 and other carboxylato compounds there is no Cr-Cr bond.’4aThis, however, has been conclusively r e f ~ t e d ’ ~by~ a, ’later ~ ab initio calculation that suggests that the Cr-Cr bond should have its maximum strength at a distance of ca. 2.40 Later studies on the carboxylate series Cr2(02CR)4L2, concerned with examining the factors contributing to the Cr-Cr bond, have shown that the variations of the Cr-Cr bond distance in this series cannot be accounted for solely by postulating linear and independent dependencies on the inductive effect of the R groups and the distance of the axial ligands.” Recent investigations designed to examine A;73’09’1

0 1980 American Chemical Society